- From: Graham Klyne <GK@ninebynine.org>
- Date: Thu, 15 Feb 2001 09:52:48 +0000
- To: "Bill de hOra" <bill@dehora.fsnet.co.uk>
- Cc: RDF interest group <www-rdf-interest@w3.org>
At 12:58 AM 2/15/01 +0000, you wrote: >Ok. I don't see the value of allowing a machine to determine two data URIs are >syntactically identical, where they could do the same with string literals >anyway. With literals it's explicit that they are semantically distinct and >can't be safely unified. With data: URIs it's explicit that they are >semantically distinct and can't be safely unified. What extra do you get from >data URIs, other than a mimetype (which seems questionable)? What you get, IMO, is a simplification of the RDF model. >Though mind you that does raise a point. Since data URIs don't refer to >resources, do the set of resources need to explicitly exclude the data URI as >being possible identifiers? Has this come up before? I think data URIs _do_ refer to resources. Very primitive resources that map the URI to a specified item of data (and some associated MIME labelling). #g ------------ Graham Klyne GK@NineByNine.org
Received on Thursday, 15 February 2001 05:11:49 UTC